#8413 CUDA GPU acceleration unavailable. Verify NVIDIA GPU & CUDA installation (Help -> CUDA Help).

This is because CUDA acceleration can be used only when an NVIDIA GPU, an up-to-date driver, CUDA 12.8, and cuDNN 9.x are all present and correctly configured; if any of these is missing, the software falls back to CPU.

Troubleshooting steps:

1. Confirm the GPU is NVIDIA. Integrated or discrete GPUs from AMD or Intel cannot use CUDA at all; NVIDIA is required.
2. Update the NVIDIA driver to the latest version from the official site. CUDA 12.8 has a minimum driver version requirement; an outdated driver is the most common cause.
3. Confirm CUDA 12.8 is installed and that nvcc -V works in the command line.
4. Install cuDNN 9.x, then add both its bin and lib directories to the system Path environment variable, and reboot so the change takes effect.
5. Change the software path. The current directory D:/SOFT TAO BAO 10-10-26/电脑手机屏幕翻译/电脑/win-pyvideotrans-v4.17 contains spaces and Chinese characters, which easily causes library loading failures. Move it to a short path with only English letters and digits, for example D:/pyVideoTrans.
6. After the environment is ready, restart the software and check CUDA Acceleration on the main interface.

Note: if the environment still fails, the software will keep running on CPU; startup will not fail, but processing will be much slower. If you only see this prompt at startup, you can ignore it and run on CPU for now.

Please refer to the related documentation:

• Configuring CUDA and cuDNN on Windows